Alprazolam (Xanax): Independent Evidence on Panic, Anxiety & Withdrawal

by
Key takeaways
  • Alprazolam (Xanax) is a fast-acting benzodiazepine. In the US it is licensed for the acute treatment of generalised anxiety disorder (GAD) and for panic disorder (FDA Xanax label). In the UK it holds a licence only for short-term treatment of severe anxiety, for a maximum of 2–4 weeks (UK SmPC).
  • Although it is licensed in the UK, alprazolam cannot be prescribed on the NHS (Corkery et al., J Psychopharmacol 2022; NHS Borders ADP briefing 2018).
  • In the short term it works, especially for panic. An independent Cochrane network meta-analysis ranked alprazolam among the most effective and best-tolerated drugs for panic disorder. However, it rated the benzodiazepine trials as low quality (Guaiana et al., Cochrane 2023).
  • Guidelines advise against ongoing use. NICE says benzodiazepines "should not be prescribed" for panic disorder, and for GAD only "as a short-term measure during crises" (NICE CG113). The WHO makes a strong recommendation against them for GAD and panic, apart from emergency use for 3–7 days at most (WHO mhGAP 2023).
  • Its FDA boxed warning covers three risks: deaths when taken with opioids, abuse and addiction, and withdrawal reactions that can be life-threatening. In 2019, alprazolam made up 38% of the estimated 92 million US benzodiazepine prescriptions, the largest share of any benzodiazepine (FDA 2020).
  • Counterfeit "Xanax" is a documented cause of death. In 2021 US overdose data, 17.0% of deaths with evidence of counterfeit pills involved counterfeit alprazolam only, and 93.0% of all counterfeit-pill deaths involved illicit fentanyls (CDC MMWR 2023). Scotland registered 370 alprazolam-related deaths from 2004 to 2020 (Corkery 2022).
  • Evidence grade: Moderate for short-term relief of panic and anxiety. Insufficient for long-term benefit. Risk for dependence, withdrawal and misuse.

Independent evidence review · Prescription medicine

Alprazolam calms anxiety and panic within hours, and short-term trials consistently beat placebo. That speed is also the problem: it is the most-prescribed benzodiazepine in the US, one of the most misused, and one of the hardest to stop. UK and WHO guidance treat it as a crisis tool for days to a few weeks, not a treatment for anxiety disorders. Most of the efficacy data come from short trials run in the 1980s and 1990s for the drug's makers. Independent Cochrane reviewers agree that those trials show a benefit, but they rate the quality as low and note that none measured long-term outcomes or dependence (Breilmann et al., Cochrane 2019).

Best evidence for short-term relief of panic disorder and severe anxiety
Main risks dependence, withdrawal seizures, fatal breathing suppression with opioids or alcohol, misuse, counterfeit pills
Key rule shortest possible course; never stop suddenly; never combine with alcohol or opioids; never take pills from a non-pharmacy source
Safety first
Alprazolam is a prescription-only controlled drug. Do not start it, stop it or change the dose without the prescriber who manages your treatment. Stopping suddenly after regular use can cause severe withdrawal, including seizures, which can be life-threatening (FDA boxed warning). Taking it with opioids (including codeine and tramadol), alcohol or other sedatives can cause profound sedation, slowed breathing, coma and death. Do not drive or use machinery until you know how it affects you (UK SmPC 4.7). Benzodiazepines should not be used alone to treat depression, because they may increase the risk of suicide. If you have thoughts of harming yourself, seek urgent help now. In the UK, call NHS 111 or 999. If someone is very drowsy, breathing slowly or cannot be woken, call 999 (or your local emergency number). Tablets sold as "Xanax" outside pharmacies may contain fentanyl or unpredictable designer benzodiazepines (CDC).

Table of contents

Evidence summary

Claim Evidence Source Funding / conflict Strength
Benzodiazepines beat placebo for panic disorder in the short term 24 studies, 4,233 participants. Response RR 1.65 (95% CI 1.39–1.96), NNTB 4. Remission RR 1.61. The overall methodological quality was poor, and the evidence was rated low quality. Breilmann et al., Cochrane 2019 Cochrane team (Germany/Italy). Nine authors declared no conflicts. One declared lecture fees from Eli Lilly, Janssen, MSD, Pfizer and others. Moderate (low-quality trials)
Alprazolam ranks among the most effective and best-tolerated panic drugs 70 trials. For response, diazepam, alprazolam and clonazepam ranked as most effective. Alprazolam and diazepam had lower dropout rates than placebo. Only clonazepam and alprazolam showed a strong reduction in panic-attack frequency. The benzodiazepine comparisons were low quality. Guaiana et al., Cochrane 2023 Cochrane (UK/Italy/Canada/Japan). Several authors declared fees from Otsuka, Pfizer, Eli Lilly and others. No declared ties to Xanax's makers were identified in the declarations. Moderate
Benzodiazepines are effective for GAD but poorly tolerated 89 trials, 25,441 patients. "Paroxetine and benzodiazepines were effective but also poorly tolerated when compared with placebo." Slee et al., Lancet 2019 "No funding was received" (UCL, UK). Moderate
Alprazolam as effective as antidepressants for panic Response RR 0.99 (95% CI 0.67–1.47), but based on only 2 studies with 215 participants. Low-quality evidence. Bighelli et al., Cochrane 2016 Cochrane (Italy). One author was an expert witness for Accord Healthcare; another declared pharma lecture fees. Insufficient
Long-term benefit beyond 8 weeks 8 studies (N = 1,228) lasting 13 weeks or more. No significant HAM-A difference between benzodiazepines and placebo. Benzodiazepines were no different from antidepressants in those who had already responded. Shinfuku et al., Int Clin Psychopharmacol 2019 Japanese academic team (Keio University). The PubMed record gives no funding statement. Insufficient
Short-term GAD and panic efficacy in the licensing trials In 4-week GAD trials, Xanax beat placebo on all rating scales. In 3 panic trials of up to 10 weeks, 37% to 83% of Xanax patients had zero panic attacks, and Xanax beat placebo on this measure in all 3. FDA Xanax label §14 Trials submitted by the NDA holder (Upjohn, now Viatris). The label is FDA-reviewed, but the data are the manufacturer's. Moderate
Physical dependence and withdrawal Discontinuation symptoms in 641 patients: insomnia 29.5%, light-headedness 19.3%, anxiety 19.2%, abnormal involuntary movement 17.3%. Withdrawal seizures reported. Protracted withdrawal can last "weeks to more than 12 months". FDA label §5.3, §6.1, §9.3 Regulator-required safety text Risk
Deaths with opioids and illicit supply Scotland: 370 alprazolam-related deaths from 2004 to 2020. Opiates or opioids were implicated in 94.8%, and alprazolam alone in two deaths. Corkery et al., 2022 University of Hertfordshire (UK). Two authors have served on the UK Advisory Council on the Misuse of Drugs; one declared fees from Pfizer and other companies. Risk

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic or Cochrane team with no declared commercial funding for the work; 3 = independent authors analysing mostly manufacturer-run trials, or minor declared industry ties; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Source Who funds it Country Independence tier (1–4) Credibility (A–D) Why it would tell the truth / residual bias
NICE CG113 UK Department of Health and Social Care (grant-in-aid) UK 1 A Must justify NHS spending and has no product to sell. The panic recommendation dates from 2004, and the committee's conflict-of-interest registers were not checked here.
WHO mhGAP guideline 2023 World Health Organization International 1 A Uses GRADE and states its certainty openly (low). It is written for low-resource settings, which may push it towards caution with dependence-forming drugs.
FDA safety communications and Xanax label US government. In FY2025, 77% of the cost of human drug review came from industry user fees (FDA PDUFA report) US 1 (safety text) / 4 (efficacy data) A for warnings The warnings are legally binding and cut against sales. The efficacy sections summarise the manufacturer's own trials.
UK SmPC (Xanax) Written by Viatris, approved by the MHRA UK 1–4 A for restrictions The 2–4 week limit and dependence warnings are regulator-imposed. The rest is the company's text.
Breilmann et al., Cochrane 2019 Cochrane (commercial sponsorship not allowed) Germany / Italy 2 A Nine of ten authors declared no conflicts, and it openly downgrades its own evidence. It can only pool the trials that exist, most of which were industry-run.
Guaiana et al., Cochrane 2023 Cochrane; one author NIHR-funded International 2–3 A− A network meta-analysis that includes 70 trials. Some authors declared unrelated industry fees. It measured only acute efficacy and dropout, not dependence.
Slee et al., Lancet 2019 No funding UK 2 A− Searched regulator and company registries to reduce publication bias. Its benzodiazepine findings are at class level only.
AGS Beers Criteria 2023 American Geriatrics Society; "no sponsor" US 2 A− A professional society whose remit is harm reduction in older people. Some panellists have consulting ties, which are disclosed.
CDC MMWR 2023 US government US 1 A Uses death-certificate and toxicology surveillance data. It covers only 35 US jurisdictions, and "evidence of counterfeit pill use" depends on what investigators recorded.
Corkery et al., J Psychopharmacol 2022 Academic (University of Hertfordshire) UK 2 B+ Based on official Scottish mortality registers. The authors have public-advisory roles, and one declared industry fees.
Viatris 10-K (FY2025) Viatris US 4 A for its own sales figures Filed under legal penalty for misstatement. It is the company's own account.

What alprazolam is

Alprazolam is a triazolo-benzodiazepine, a "1,4 benzodiazepine" in the FDA's description (FDA label §12.1). It is absorbed quickly: peak blood levels come 1 to 2 hours after a dose. It is cleared fairly quickly too, with a mean half-life of about 11.2 hours (range 6.3 to 26.9) in healthy adults. This puts it among the short-acting benzodiazepines, unlike long-acting diazepam.

  • Brand names: Xanax (tablets) and Xanax XR (extended-release, US). The FDA's benzodiazepine list names both (FDA 2020).
  • First approvals: The US FDA approved Xanax tablets (NDA 018276, sponsor Upjohn) on 16 October 1981 (Drugs@FDA). The UK first authorised Xanax on 27 August 1982 (UK SmPC section 9).
  • Originator: The Upjohn Company (US) was the original NDA sponsor. US Xanax packs now state "Distributed by Pharmacia & Upjohn Co Division of Pfizer Inc". Adverse-event reports go to Viatris (FDA label).
  • Current owner: Viatris Inc. (Canonsburg, Pennsylvania, US) was formed on 16 November 2020 when Mylan combined with Pfizer's Upjohn business. Xanax is one of its reported brands (Viatris 10-K). The UK licence holder is Viatris Products Limited, Potters Bar (PL 46302/0328) (UK SmPC).
  • Generic status: Off-patent. openFDA lists 11 generic (ANDA) alprazolam applications (openFDA query).
  • US status: Prescription-only, Schedule IV controlled substance (FDA label §9.1).
  • UK status: Licensed prescription-only medicine. Only the 250 microgram and 500 microgram Xanax tablets appear on the UK medicines compendium (eMC listing). It is a Class C drug under the Misuse of Drugs Act 1971 (Schedule 2, Part III) and a Schedule 4 Part I controlled drug under the Misuse of Drugs Regulations 2001 (Schedule 4). It "cannot be prescribed on the National Health Service" (Corkery 2022). An NHS Scotland drugs partnership says the same: "It is not prescribed under NHS" (NHS Borders ADP 2018). We could not open the NHS Drug Tariff listing itself to confirm the exact entry. The NHS website has no alprazolam medicines page.

How it works

According to the FDA label, alprazolam binds to the benzodiazepine site of the GABAA receptor in the brain and "enhances GABA-mediated synaptic inhibition" (FDA label §12.1). GABA is the brain's main inhibitory messenger. Strengthening its effect reduces anxiety, but it also causes sedation, poorer coordination, memory lapses and, at high doses or with other depressants, slowed breathing.

Pharmacokinetics that matter in practice (FDA label §12.3):

  • Metabolism: Mainly by the liver enzyme CYP3A4. This drives its most important drug interactions. Ketoconazole raised alprazolam exposure (AUC) 3.98-fold, itraconazole 2.66-fold, nefazodone 1.98-fold, fluvoxamine 1.96-fold and erythromycin 1.61-fold.
  • Half-life varies by group:
    • Healthy older adults: 16.3 hours, against 11.0 hours in younger adults
    • Obese adults: 21.8 hours
    • Alcoholic liver disease: 19.7 hours
    • Asian people: peak concentrations about 15% higher and half-life about 25% longer than in Caucasian people
  • Smoking: Can lower alprazolam levels by up to 50%.
  • Interdose symptoms: Because it is short-acting, early-morning anxiety and anxiety emerging between doses have been reported in people with panic disorder (FDA label §5.3). The UK SmPC notes that withdrawal symptoms may occur within the dosing interval with short-acting benzodiazepines, especially at high doses (UK SmPC 4.4).

What it is prescribed for

Licensed indications differ by country

Condition US (FDA, Xanax) UK (Xanax SmPC)
Generalised anxiety disorder"Acute treatment" in adultsCovered by the general anxiety indication below
Anxiety (general)—"Short-term symptomatic treatment of anxiety in adults", only when it is "severe, disabling or subjecting the individual to extreme distress"; maximum 2–4 weeks
Panic disorder, with or without agoraphobiaAdultsNot licensed
Children and adolescentsSafety and effectiveness not establishedNot recommended under 18
NHS availability—Not prescribable on the NHS

Sources: FDA label §1, §8.4; UK SmPC 4.1, 4.4; Corkery 2022.

Where guidelines place it

  • GAD (NICE CG113, 1.2.26): "Do not offer a benzodiazepine for the treatment of GAD in primary or secondary care except as a short-term measure during crises." The first-choice drug is an SSRI (sertraline first), alongside or after psychological therapy (NICE CG113). In the UK, therefore, alprazolam is at most a crisis-only option.
  • Panic disorder (NICE CG113, 1.3.20): "Benzodiazepines are associated with a less good outcome in the long term and should not be prescribed for the treatment of individuals with panic disorder" (NICE CG113). This is not recommended, even though panic is alprazolam's main US indication.
  • WHO mhGAP (2023), recommendation ANX6: "Benzodiazepines are not recommended for the treatment of adults with generalized anxiety disorder (GAD) and/or panic disorder." They may be considered for emergency management of acute, severe anxiety, but only as a short-term (3–7 days maximum) measure. This is a strong recommendation based on low-certainty evidence (WHO mhGAP).
  • NHS patient guidance: For GAD, a GP "will usually advise you to try talking therapies before they prescribe medicine" (NHS).
  • Older adults (AGS Beers 2023): Avoid benzodiazepines, including alprazolam. The criteria note they "may be appropriate" for a few uses, such as severe generalised anxiety disorder and alcohol or benzodiazepine withdrawal (AGS Beers 2023).

What works and what does not

Claimed benefit Verdict Evidence Key caveat
Short-term relief of panic disorder WORKS Benzodiazepines vs placebo: response RR 1.65, NNTB 4 (Breilmann 2019). Alprazolam is among the top-ranked drugs for response and panic frequency (Guaiana 2023). Low-quality, short trials with "probable publication bias". None measured dependence.
Short-term relief of GAD symptoms WORKS Better than placebo in 4-week trials (FDA label). As a class, benzodiazepines were effective but poorly tolerated (Slee 2019). NICE and WHO both advise against it except in crises.
Better than antidepressants for panic INSUFFICIENT No difference in response (RR 0.99), but based on only 2 trials (Bighelli 2016). No difference between classes in the network meta-analysis (Guaiana 2023). Antidepressants carry no addiction risk, which is why guidelines prefer them.
Long-term anxiety control (months to years) INSUFFICIENT No significant HAM-A difference vs placebo beyond 13 weeks, from 8 small studies (Shinfuku 2019). The FDA label says the necessary duration of treatment for panic "is unknown". Tolerance to the calming effect may develop, but "little tolerance develops to the amnestic reactions and other cognitive impairments" (FDA §9.3).
Treating depression NOT RECOMMENDED The UK SmPC says benzodiazepines "should not be prescribed alone to treat depression as they may precipitate or increase the risk of suicide". Hypomania and mania have been reported with alprazolam in depression (UK SmPC 4.4). —
Sleep problems NOT LICENSED Neither the US nor the UK licenses alprazolam for insomnia. The UK SmPC notes some loss of hypnotic effect after repeated use for a few weeks.

Benefits by claim

Panic disorder

Panic disorder is where alprazolam's evidence is strongest, and it is the use for which it became famous in the US. The FDA label describes three short-term, placebo-controlled trials lasting up to 10 weeks. Two used average doses of 5 to 6 mg a day, and the third used fixed doses of 2 mg and 6 mg. In all three, Xanax beat placebo on the proportion of patients with zero panic attacks (the range across Xanax groups was 37% to 83%) and on global improvement. In two of the three, it also reduced weekly panic attacks and phobia ratings (FDA label §14.2). These are the manufacturer's trials as summarised by the regulator. The label does not give the placebo rates for the zero-attack outcome.

The independent reviews broadly agree that it works in the short term:

  • Breilmann et al. (Cochrane 2019): Pooled 24 double-blind trials of benzodiazepines against placebo, with 4,233 participants.
    • Response: RR 1.65 (95% CI 1.39 to 1.96), NNTB 4 (95% CI 3 to 7).
    • Remission: RR 1.61.
    • Dropout was lower on benzodiazepines (RR 0.50).
    • However, more people stopped because of adverse effects (RR 1.58, 95% CI 1.16 to 2.15).
    • More people had at least one adverse effect (RR 1.18).
    The authors rated the evidence as low quality and the included studies as methodologically "poor". They flagged "possible unmasking of allocated treatments, high dropout rates, and probable publication bias". They also noted that none of the studies examined long-term efficacy or the risks of dependency and withdrawal (Breilmann 2019).
  • Guaiana et al. (Cochrane 2023): A network meta-analysis of 70 trials.
    • Across 48 RCTs (N = 10,118), "diazepam, alprazolam and clonazepam ranking as the most effective" for response.
    • Across 64 RCTs, alprazolam and diazepam had lower dropout rates than placebo and ranked as the most tolerated drugs.
    • Only clonazepam and alprazolam showed a strong reduction in panic-attack frequency.
    There were two important qualifications. Between drug classes, "we found no difference" in response. The quality of the studies comparing benzodiazepines with placebo and with antidepressants "was low" (Guaiana 2023).

What this means clinically. An NNTB of 4 is a large short-term effect for a psychiatric drug. Benzodiazepines also have low dropout rates in trials. The reason this does not translate into a guideline recommendation is time. Panic disorder is often long-lasting, and the trials stop at weeks. NICE judged that benzodiazepines are "associated with a less good outcome in the long term" (NICE CG113). The FDA label itself says tapering off after extended freedom from panic "may often be difficult to accomplish without recurrence of symptoms and/or the manifestation of withdrawal phenomena" (FDA label §2.2).

Generalised anxiety disorder

The FDA label reports that, in 4-week placebo-controlled trials at doses up to 4 mg a day, Xanax was "significantly better than placebo at each of the evaluation periods" on the Hamilton Anxiety scale and other measures. No effect sizes are given (FDA label §14.1). The largest independent network meta-analysis of GAD drugs (89 trials, 25,441 patients, no funding) found that "paroxetine and benzodiazepines were effective but also poorly tolerated when compared with placebo" (Slee et al., Lancet 2019). The abstract does not report an alprazolam-specific effect size. The WHO reviewed the same body of evidence and still recommended against benzodiazepines for GAD, "since the GDG concluded that the risks of the intervention outweighed the benefits" (WHO mhGAP).

Placebo response and how fast it works

Placebo response in anxiety trials is substantial. In the GAD trials behind the US label, for example, 22% of placebo patients reported drowsiness and 19% light-headedness (FDA Table 1). Sedation is easy to notice, which is one reason Cochrane reviewers worry that blinding broke down in benzodiazepine trials. Alprazolam's peak blood level is reached 1 to 2 hours after a dose (FDA §12.3). Breilmann and colleagues write that benzodiazepines remain widely used "probably because of their rapid onset of action" (Breilmann 2019).

Risks and all side effects

The FDA boxed warning, summarised

Since the FDA's class-wide action in September 2020, the Xanax label carries a boxed warning covering three risks (FDA label; FDA 2020):

  • Opioids: Using benzodiazepines with opioids "may result in profound sedation, respiratory depression, coma, and death". Combined prescribing should be reserved for patients with no adequate alternative. This part was first added in August 2016. The FDA then reported that emergency visits involving non-medical use of both drug types rose from 11 to 34.2 per 100,000 population between 2004 and 2011 (FDA 2016, archived).
  • Abuse, misuse and addiction: These "can lead to overdose or death". Prescribers should assess each patient's risk before and throughout treatment.
  • Dependence and withdrawal: Abrupt stopping or rapid dose reduction "may precipitate acute withdrawal reactions, which can be life-threatening". A gradual taper is required.

The FDA's 2020 review also reported the scale of the problem. US benzodiazepine-involved overdose deaths rose from 1,298 in 2010 to 11,537 in 2017. From 2013 to 2017, 55% of these deaths also involved prescription opioids. In 2018, an estimated 50% of patients dispensed oral benzodiazepines received them for two months or longer (FDA 2020). In the UK, the MHRA told prescribers in March 2020 to "only prescribe benzodiazepines … and opioids together if there is no alternative" (MHRA Drug Safety Update).

Common side effects (licensing trials)

Side effect / concern Frequency (Xanax vs placebo) Context Practical note Source
Drowsiness41% vs 22% (GAD); 77% vs 43% (panic)Panic trials used up to 10 mg/dayDo not drive if affected.FDA Tables 1–2
Fatigue and tiredness49% vs 42% (panic)——FDA Table 2
Impaired coordination40% vs 18% (panic)—A falls risk, especially in older people.FDA Table 2
Memory impairment; cognitive disorder33% vs 22%; 29% vs 21% (panic)—Anterograde amnesia is possible. The UK SmPC advises ensuring 7–8 hours of uninterrupted sleep.FDA Table 2; SmPC 4.4
Slurred speech (dysarthria)23% vs 6% (panic)——FDA Table 2
Light-headedness21% vs 19% (GAD)——FDA Table 1
Dry mouth15% vs 13% (GAD)——FDA Table 1
Constipation26% vs 15% (panic)——FDA Table 2
Weight and appetite changesIncreased appetite 33% vs 23%; weight gain 27% vs 18%; weight loss 23% vs 17% (panic)—It can go either way.FDA Table 2
Libido and sexual functionDecreased libido 14% vs 8%; sexual dysfunction 7% vs 4% (panic)Increased libido is also listed—FDA Table 2
Low blood pressure5% vs 2% (GAD)——FDA Table 1

Serious risks

Risk What is known Who is most at risk Practical note Source
Fatal respiratory depression with opioids, alcohol or other sedativesBoxed warning. In Scotland, 94.8% of alprazolam-related deaths involved opiates or opioids, and 64.2% involved combinations of CNS depressants.People taking opioids, gabapentinoids or alcohol, or using illicit drugsNever combine without explicit specialist agreement. Naloxone reverses opioids, not benzodiazepines.FDA; Corkery 2022; NHS Borders
Withdrawal seizures and severe withdrawalLife-threatening reactions include seizures, delirium, psychosis and suicidality. In 641 patients, discontinuation symptoms included insomnia (29.5%), anxiety (19.2%), nausea or vomiting (16.5%), sweating (14.4%) and tachycardia (12.2%).Doses above 4 mg a day and longer useTaper gradually under supervision.FDA §5.3, Table 3, §9.3
Protracted withdrawal syndromeAnxiety, cognitive problems, insomnia, tinnitus and tingling lasting "weeks to more than 12 months"Long-term usersThis can be hard to tell apart from the original anxiety returning.FDA §9.3
Dependence even at therapeutic doses"Even after relatively short-term use at doses of ≤4 mg/day, there is some risk of dependence." The UK SmPC says it can occur "at therapeutic doses and/or in patients with no individualised risk factor".Higher doses, use beyond 12 weeks, or a history of substance use or mental health disorderAgree an exit plan before starting.FDA §5.3; SmPC 4.2, 4.4
Abuse and diversionThe UK SmPC warns that alprazolam "may be subject to diversion". In 2019 it was the most-dispensed US benzodiazepine, at 38% of prescriptions.People with addictive disorders; household membersStore securely and dispose of unused tablets.SmPC 4.4; FDA 2020
Counterfeit "Xanax"Counterfeit pills are "frequently" made to look like oxycodone or alprazolam. In 2021, 2,437 US overdose deaths (4.4% of those studied) showed evidence of counterfeit pill use. Of these, 17.0% involved counterfeit alprazolam only, and 93.0% involved illicit fentanyls.Anyone buying outside a pharmacyNo pill from an online or street source can be assumed to be alprazolam.CDC MMWR 2023
Falls, fractures, delirium and car crashes in older adults"All benzodiazepines increase the risk of cognitive impairment, delirium, falls, fractures, and motor vehicle crashes in older adults." Shorter-acting ones "are not safer".Age 65 and overBeers 2023: avoid (strong recommendation).AGS Beers 2023
Paradoxical reactionsAgitation, rage, aggression and hallucinations are reported rarely. After-effects in illicit use include increased hostility and aggression.Children, older adults, people taking other CNS drugsStop and seek advice.FDA §6.1; SmPC 4.4; NHS Borders
Worsening depression, mania or hypomaniaMania and hypomania reported in patients with depressionPeople with depression or bipolar disorderPrescribe the smallest feasible supply.FDA §5.6; SmPC 4.4
Interdose rebound anxietyEarly-morning anxiety or anxiety between dosesPanic disorder patients on short-acting dosingReport it to your prescriber rather than adding doses.FDA §5.3
Neonatal sedation and withdrawalFloppy infant syndrome, breathing problems and withdrawal after late-pregnancy usePregnancy, especially the third trimesterDiscuss with a prescriber before or during pregnancy.FDA §8.1; SmPC 4.6
Angioedema and hypersensitivityAngioedema reported—Swelling of the face or throat is an emergency.FDA §4

Overdose. The UK SmPC says benzodiazepine overdose "should not present a threat to life unless combined with other CNS depressants (including alcohol)" (SmPC 4.9). The Scottish data show why that caveat matters. Of 370 alprazolam-related deaths, only two involved alprazolam alone (Corkery 2022).

The UK illicit-supply problem

Because alprazolam cannot be prescribed on the NHS, most alprazolam in UK circulation is illicit. An NHS Scotland drugs partnership warned in 2018 that its "availability is increased through illegal supply via online sources". It described 2 mg "bars" containing 8 times the smallest active dose, and noted that the drug is used as a "comedown aid" after stimulants (NHS Borders ADP 2018). Scotland registered 370 deaths mentioning alprazolam from 2004 to 2020, and 366 of them occurred from 2015 onwards. Most of those who died were men (77.1%), with a mean age of 39. The researchers attribute the fall in 2019–2020 partly to people switching to designer benzodiazepines (Corkery 2022).

All interactions

Interacts with Examples Severity Mechanism Action
OpioidsMorphine, oxycodone, codeine, dihydrocodeine, tramadol, fentanyl, methadone, buprenorphineBoxed warningAdditive respiratory depression (GABAA plus mu-opioid receptors)Only if there is no alternative, at the lowest doses and shortest duration, with monitoring.
Strong CYP3A inhibitorsKetoconazole, itraconazole, clarithromycinContraindicated (US)Alprazolam exposure up to 3.98-fold higherDo not combine (ritonavir is handled separately; see below).
AlcoholAny amountAvoidAdditive CNS depressionThe UK SmPC says combined intake "is not recommended". The FDA says "Do not drink alcohol with benzodiazepines".
Other CNS depressantsOther benzodiazepines, Z-drugs, gabapentin and pregabalin, sedating antihistamines, antipsychotics, antiepileptics, anaesthetics, cannabisHigh cautionAdditive sedation and respiratory depressionGabapentinoids were implicated in 42.9% of Scottish alprazolam-related deaths (Corkery 2022).
Moderate or weak CYP3A inhibitorsFluvoxamine (1.96-fold), nefazodone (1.98-fold), erythromycin (1.61-fold), cimetidine (Cmax +82%), fluoxetine (Cmax +46%), oral contraceptives (Cmax +18%)Avoid or reduce doseSlower clearance of alprazolamThe prescriber may lower the dose.
RitonavirHIV and COVID-19 regimens that contain ritonavirDose adjustmentShort-term inhibition, then induction after 10–14 daysHalve the alprazolam dose when starting both together.
CYP3A inducersCarbamazepine, phenytoinCautionCarbamazepine shortened the alprazolam half-life from 17.1 to 7.7 hoursReduced effect is possible.
DigoxinDigoxinMonitorRaised digoxin levels, especially over age 65Check digoxin levels.
Imipramine, desipramineTricyclic antidepressantsMonitorTricyclic levels rose by 31% and 20% respectively—
FlumazenilBenzodiazepine antagonist (hospital use)SpecialistCan precipitate acute withdrawal, including seizuresFor hospital teams only.
SmokingTobaccoNoteAlprazolam levels up to 50% lowerTell your prescriber if you start or stop smoking.

Sources: FDA label §2.6, §4, §7, §12.3; UK SmPC 4.5; Corkery 2022. The FDA label notes that a single alprazolam dose with steady-state sertraline caused no clinically significant pharmacokinetic change.

Who should avoid alprazolam

  • Contraindicated (UK): hypersensitivity to benzodiazepines, myasthenia gravis, severe respiratory insufficiency, sleep apnoea syndrome and severe hepatic insufficiency (UK SmPC 4.3).
  • Contraindicated (US): hypersensitivity (angioedema has been reported) and use with strong CYP3A inhibitors other than ritonavir (FDA §4).
  • People taking opioids, or with a current or past substance-use disorder, including alcohol misuse. The UK SmPC says to use it with "extreme caution" if there is a history of alcohol or drug abuse.
  • People with depression or suicidal thoughts: Do not use it alone. Limit the supply.
  • Older adults: The AGS Beers Criteria 2023 say avoid (strong recommendation). If used, the US starting dose is 0.25 mg two or three times a day, because clearance is reduced (Beers 2023; FDA §2.4).
  • Liver disease: The half-life averaged 19.7 hours in alcoholic liver disease. A lower starting dose is required, and use is contraindicated in severe insufficiency in the UK.
  • Kidney impairment: The UK SmPC advises caution.
  • Pregnancy: Observational studies "do not report a clear association" with major birth defects. Some early case-control studies found a twofold increased risk of oral clefts (UK SmPC). Late-pregnancy use can cause neonatal sedation and withdrawal (FDA §8.1; SmPC 4.6).
  • Breastfeeding: "Breastfeeding not recommended" during treatment (FDA §8.2).
  • Under 18s: Not recommended. Safety and efficacy have not been established.
  • People with lung disease, such as COPD, because of the risk of respiratory depression.

Dosage and how to take it

Your prescriber sets the dose. The figures below are the official licensed adult ranges, not personal advice.

Use UK licence (Xanax SmPC) US licence (FDA Xanax label)
Anxiety / GAD0.25–0.5 mg three times daily, increasing if required to a total of 3 mg daily. Maximum treatment 2–4 weeks.Start at 0.25–0.5 mg three times daily. Increase every 3–4 days if needed, to a maximum of 4 mg/day in divided doses.
Panic disorderNot licensedStart at 0.5 mg three times daily. Increase by no more than 1 mg/day every 3–4 days. Trials used 1–10 mg/day, with a mean of about 5–6 mg/day.
Older or debilitated adults0.25 mg two to three times daily0.25 mg two or three times daily
Liver impairmentCaution in mild to moderate impairment; contraindicated in severeStart at 0.25 mg two or three times daily

Sources: UK SmPC 4.2; FDA label §2.

  • How long before it works: Peak blood levels come 1–2 hours after a dose (FDA §12.3). This is much faster than antidepressants.
  • How long to take it: In the UK, "the lowest possible effective dose, for the shortest possible time and for a maximum of 2-4 weeks". The prescriber should agree a plan for ending treatment before starting (UK SmPC 4.2).
  • How to stop: Always taper under supervision. The US label advises reducing by "no more than 0.5 mg every 3 days", and says "some patients may require an even slower dosage reduction". If withdrawal symptoms appear, the taper may be paused or the dose stepped back up before slowing down. "Some patients may prove resistant to all discontinuation regimens" (FDA label §2.3). In a controlled postmarketing study, a slower taper did not change how many people reached zero, but it reduced withdrawal symptoms.
  • UK withdrawal guidance: NICE advises "a slow, stepwise rate of reduction proportionate to the existing dose, so that decrements become smaller as the dose is lowered". For a benzodiazepine with a short half-life, such as alprazolam, it says prescribers should consider switching to one with a longer half-life (NICE NG215, 1.5.11–1.5.13).
  • If a higher dose is needed, the UK SmPC says the evening dose should be increased before the daytime doses.

Follow the money: who makes it and who funded the evidence

Ownership chain

  • 1981: The Upjohn Company (Kalamazoo, Michigan, US) gains FDA approval of Xanax under NDA 018276 (Drugs@FDA). The UK tablets are still marked "Upjohn 29" (UK SmPC).
  • Pharmacia & Upjohn, then Pfizer: US packs read "Distributed by Pharmacia & Upjohn Co Division of Pfizer Inc, NY" (FDA label). We did not trace the dates of the Pharmacia and Pfizer mergers to a primary source in this review.
  • 2020 onwards: Viatris Inc. (Canonsburg, Pennsylvania, US) was formed on 16 November 2020 by "Mylan combining with Pfizer's Upjohn Business" (Viatris 10-K FY2025). Viatris Products Limited holds the UK licence.

Revenue

Viatris reports Xanax net sales of $139.9 million in 2025, $145.0 million in 2024 and $154.8 million in 2023. That is a small, slowly declining line in a company with total 2025 net sales of $14,250.4 million (Viatris 10-K FY2025). Most alprazolam is now dispensed as generics: openFDA lists 11 ANDA holders. So the brand owner has limited financial stake in today's prescribing volume.

Who funded the evidence

  • Licensing trials: The GAD and panic trials summarised in the FDA label were submitted by the NDA holder. The label does not name each trial's sponsor, and we did not verify individual trial funding. Treat these as manufacturer evidence.
  • Independent syntheses:
    • The Cochrane panic reviews were produced under Cochrane's policy that reviews "cannot be funded or conducted by commercial sponsors" (Cochrane policy). Individual authors' pharma ties, none to Viatris or Pfizer's Upjohn unit specifically except a Pfizer lecture fee, are disclosed in the table above.
    • Slee 2019 received no funding.
    • The WHO and NICE guidelines are publicly funded.
  • Regulators: 77% of FDA human-drug review costs in FY2025 came from industry user fees (FDA PDUFA report). The boxed warnings nonetheless run against commercial interest.
  • Harm data: The CDC (US government) and the University of Hertfordshire analysis of official Scottish registers do not depend on manufacturer funding.

Integrity events: We found no sourced regulatory enforcement action specific to Xanax marketing in the documents fetched for this review. We do not infer one.

For wider, evidence-graded context, see these Pure City Research guides:

Sibling medicine reviews:

Frequently asked questions

How long does Xanax take to work?

Alprazolam reaches peak blood levels 1 to 2 hours after a dose (FDA label §12.3). That speed is why it is attractive in a crisis. It is also why guidelines limit it: the relief is short-lived, and regular use can lead to dependence.

Can I get alprazolam on the NHS?

No. Xanax holds a UK marketing authorisation for short-term severe anxiety (UK SmPC), but alprazolam "cannot be prescribed on the National Health Service" (Corkery 2022). NICE advises against benzodiazepines for panic disorder and allows them for GAD only briefly in a crisis (NICE CG113). Talk to your GP about NHS options such as talking therapies and SSRIs.

Can you drink alcohol on Xanax?

No. The FDA says "Do not drink alcohol with benzodiazepines" (FDA 2020), and the UK SmPC says combined intake "is not recommended". Alcohol adds to the sedation and breathing suppression, and most alprazolam-related deaths involve other depressant drugs (Corkery 2022).

Is Xanax addictive?

It can be. The FDA boxed warning covers abuse, misuse, addiction, physical dependence and withdrawal. The label states that "even after relatively short-term use at doses of ≤4 mg/day, there is some risk of dependence". The risk is greater above 4 mg a day and beyond 12 weeks (FDA §5.3).

How do I stop Xanax safely?

Only with your prescriber, using a gradual taper. Never stop suddenly after regular use, because withdrawal seizures can be life-threatening. The US label suggests reducing by no more than 0.5 mg every 3 days, and more slowly for some people. Some people have withdrawal symptoms lasting weeks to more than a year (FDA §2.3, §9.3).

Does Xanax cause weight gain?

In panic disorder trials, weight gain was reported by 27% on Xanax vs 18% on placebo. Weight loss was also more common, at 23% vs 17%, and so were increased appetite (33% vs 23%) and decreased appetite (28% vs 24%) (FDA Table 2). Changes can go either way.

Can I drive on alprazolam in the UK?

Not if it impairs you. It is illegal to drive while unfit because of legal or illegal drugs (GOV.UK). Alprazolam is not one of the drugs with a numerical blood limit in the 2014 regulations, which name clonazepam, diazepam, lorazepam, oxazepam and temazepam among others (SI 2014/2868). The impairment offence still applies. The UK SmPC advises patients not to drive until they know how the medicine affects them (SmPC 4.7).

Are "Xanax bars" bought online real?

You cannot tell. The CDC reports that counterfeit pills are "frequently" made to look like alprazolam, and can contain illicit fentanyls or designer benzodiazepines. In 2021, 17.0% of US overdose deaths with evidence of counterfeit pills involved counterfeit alprazolam only (CDC MMWR 2023). If someone is unresponsive, call 999 (UK) or your local emergency number.

Sources and funding notes

Last reviewed: 26 September 2026. Educational information from Pure City Research, not medical advice. Medicine decisions belong with a qualified prescriber.

Have a question — or want us to cover something?

Ask about anything on this page, or request the next deep dive: an ingredient, a supplement, or a health concern. We use published research, evidence syntheses, and regulatory guidance, with clear source links.

We store your topic, message, optional email, and this page so we can manage and reply to the request. Do not include diagnoses, medications, or other sensitive medical information. See our Privacy Policy.